The UK Civil Justice Council has recommended mandatory rules requiring lawyers and expert witnesses to disclose AI use in court documents — the most detailed framework any English-speaking jurisdiction has produced for governing AI in live litigation.
The UK Civil Justice Council has recommended mandatory rules requiring lawyers and expert witnesses to disclose when AI was used to draft court documents — the most detailed framework any English-speaking jurisdiction has produced for governing AI in live litigation.
The Council's final report draws a sharp line: factual witnesses must declare their statements were not AI-generated, while expert witnesses must disclose any AI involvement in producing their reports. Oversight responsibility falls on legal professionals and their regulators, not on blanket court restrictions. A new practice direction from the Civil Procedure Rule Committee will give the recommendations legal teeth.
Why it matters: As AI writing tools become standard across law firms and self-represented litigants alike, courts face a growing verification problem. Documents produced or refined by AI carry different reliability characteristics than purely human-authored text. The Council's framework addresses this directly — without banning AI or requiring courts to become technology auditors.
Background
The Civil Justice Council is an advisory body for the civil courts of England and Wales. Its work feeds directly into the Civil Procedure Rules, which govern how civil cases run from filing to trial. When the Council publishes recommendations, courts and practitioners take them seriously.
This report follows a period of rapid AI adoption in legal practice. In the past two years, tools like Harvey, Lexis+ AI, and general-purpose models like Claude and GPT-4 have moved from novelty to routine use at law firms of all sizes. The question is no longer whether lawyers use AI — it is whether courts know when they do.
That knowledge gap has real consequences. AI-generated text can contain confident-sounding fabrications. Expert analysis shaped by an AI model the opposing party cannot cross-examine creates fairness problems that judges haven't had to manage before. Without disclosure, courts can't assess what they're reading.
What the Report Says
The Council's core recommendations:
- Factual witness statements must include a declaration that the statement was not produced using generative AI, or disclose the extent of AI involvement if it was.
- Expert witness reports require explicit disclosure of any AI assistance in research, drafting, or analysis.
- Responsibility sits with the profession. The Council chose not to recommend that judges audit documents or build technical AI-detection infrastructure. Instead, it places the obligation on lawyers and their existing regulatory bodies — the Solicitors Regulation Authority and the Bar Standards Board — to enforce standards.
- A new practice direction from the Civil Procedure Rule Committee will formalize the disclosure requirements, giving them procedural force rather than leaving them as voluntary guidance.
The Council was explicit: its goal is not to discourage AI use — it is to make AI use visible.
Get this in your inbox.
Daily AI intelligence. Free. No spam.
Why the Expert Witness Clause Is the Sharper Edge
The expert witness provision carries more weight than the factual witness declaration. Expert reports hold evidentiary authority that witness statements do not. When an expert opines on causation, valuation, or technical risk, courts and juries treat that opinion as grounded in the expert's professional judgment.
If that judgment was substantially shaped by an AI model — one that the opposing party cannot depose, cannot examine for bias, and cannot test for the quality of its training data — the structural fairness of the proceeding is undermined. Requiring explicit disclosure forces experts and their instructing solicitors to think carefully about how AI is integrated into work product, not just whether it was used.
This is different from, say, an expert using AI to summarize documents or format a report. It is specifically targeted at AI involvement in generating the analytical content that the court will rely on.
Why This Approach Is Significant
Courts in other jurisdictions have responded to AI-generated content erratically. Some U.S. district courts have issued local rules requiring AI disclosure; others have issued outright prohibitions on AI-drafted filings. Neither approach fully solves the underlying problem.
The UK framework is more surgical. It targets the specific risks — undisclosed fabrication in witness statements, unattributed AI reasoning in expert analysis — while leaving room for AI to serve as a legitimate and useful tool in legal practice.
The placement of oversight with legal professionals rather than courts is deliberate and durable. Lawyers already carry duties of candor to the court and owe fiduciary obligations to clients. Building AI disclosure into those existing obligations is simpler and more stable than creating a new court-managed AI governance layer. If a solicitor violates the disclosure requirement, the SRA can act. That enforcement path already exists.
Impact on Legal Practice
Once the practice direction is in force, practitioners face immediate obligations:
- Case preparation workflows will need explicit AI-use tracking for any document that may enter evidence or accompany a witness statement.
- Standard engagement letters will likely be updated to address AI tool use in file management and document preparation.
- Expert witness panels at large firms and specialized agencies will need updated protocols for AI-assisted research and drafting — and those protocols will need to be client-visible.
Self-represented litigants — a growing share of civil court users — are addressed implicitly rather than explicitly. The Council's framework applies to legal professionals; non-lawyer parties are not held to the same standard. This gap is likely to generate follow-on guidance.
What to Watch
The Civil Procedure Rule Committee must turn the Council's recommendations into a formal practice direction. The timeline has not been set publicly, but practitioners should expect draft guidance within six months as courts and regulatory bodies align their policies.
More broadly, the UK framework will be watched by courts in Australia, Canada, and the United States — all jurisdictions where judicial bodies have been wrestling with the same AI disclosure question without consensus. If England and Wales sees clean adoption without procedural disruption, the model is likely to travel.
Source: UK Civil Justice Council final report on AI in court documents
Did this help you understand AI better?
Your feedback helps us write more useful content.
Get tomorrow's AI briefing
Join readers who start their day with NexChron. Free, daily, no spam.